Diabetes is traditionally classified in two broad categories: autoimmune Type 1 and obesity-related Type 2. Numerous phenotypically and etiologically distinct forms exist and are emerging, collectively termed ?atypical diabetes? (AD), that do not fit into either category. We hypothesize that AD comprises a spectrum that includes numerous forms, both known (e.g, MODY/monogenic, Latent Autoimmune Diabetes of Adults, Ketosis-Prone Diabetes) and unknown. We propose to establish the Center for Atypical Diabetes Research and Evaluation (CADRE). The goals of CADRE are to (1) identify and define comprehensively the forms of AD; and (2) establish an extensive database and repository of biosamples from AD patients. Defining the forms of AD is challenging because the etiology of AD can range from a single gene variant to a highly complex mixture of genetics, epigenetics and environmental factors. To identify patients with AD, and then to define the forms, we will use a two-pronged approach. Approach (a) will use clustering analysis of genotypic and phenotypic data from diverse groups of patients with diabetes. We have recruited 15 domestic and international population cohorts with data on more than 33,000 patients with diabetes. Further, we have recruited six diabetes study clinics (Baylor College of Medicine [BCM], University of Washington [UW], Indiana University [IU], Emory University, University of Colorado, and SUNY Downstate) who will prospectively recruit patients. The Juvenile Diabetes Research Foundation will facilitate recruitment from patient groups. The Administrative Core at BCM will work with cohorts and clinics to obtain data. The University of South Florida (USF) will filter data for AD and separate patients into phenotypically homogenous clusters that can be diagnosed based on genetic and clinical features unique to the cluster. Approach (b) will identify monogenic, oligogenic and mitochondrial forms of AD from patients enrolled in genetics registries (BCM) and a family study with genetic data (Botnia). The BCM registries include patients referred to the Undiagnosed Disease Network. All patients will undergo whole-exome and mitochondrial genome sequencing (Cores at BCM), single-nucleotide polymorphism variant analysis (Oxford), metabolomic analysis (Duke) and phenotypic analysis of C-peptide, adiponectin, proinsulin, islet autoantibodies, and serum insulin DNA levels (Cores at UW and IU). To establish the database and biorepository, we will invite patients with AD to participate in sample donation and longitudinal follow-up via the cohorts and clinics. Data and samples will be transferred to the Database and Biorepository Core at USF, with input and tracking implemented using the successful framework developed by USF for large multicenter trials. Data will be made available through the NIDDK and a CADRE website and samples will be made available through ancillary study application. Ongoing patient recruitment to expand the database and biorepository will occur through the study clinics and clinician referral through the CADRE website. Overall, CADRE will be built on a foundation of existing, successful core facilities and computational systems to create a lasting resource that will substantially advance AD research.